Advanced curriculum

Deep-dive lectures

Seven undergraduate-level lectures that broaden the assessment knowledge beyond the FastAPI track — pipeline, architectures, loss functions, ablations, and the math behind modern damage-assessment models.

L0Lecture

Project Overview · xBD + ATC-20

Master lecture: multi-architecture study on xBD with operational ATC-20 deployment. Figures, tables, equations, and the umbrella for every other module.

  • xBD dataset
  • Pipeline
  • ATC-20
L1Lecture

Vision Transformers (ViT)

Patches, self-attention, multi-head MSA, encoder block, variants (DeiT, Swin, BEiT), and how residuals defeat vanishing gradient.

  • 12 sections
  • Q/K/V math
  • Architecture diagrams
L2Lecture

Focal Loss, Dice Loss & Class Weights

Why cross-entropy fails on imbalanced disaster data, the (1−pₜ)^γ modulating factor, region-overlap losses, and combined Focal+Dice training.

  • Class imbalance
  • Worked numerics
  • PyTorch code
L3Lecture

Siamese Neural Networks

Shared encoders, difference blocks, skip connections, attention gates, deep supervision, and Siamese variants for pre/post change detection.

  • Twin networks
  • xBD pipeline
  • Deep supervision
L4Lecture

Encoder · ResNet-50

Stem, four bottleneck stages, residual shortcuts, the 7×7 deepest feature map, and the four skip-tap points that feed the U-Net decoder.

  • Backbone
  • Bottleneck zoom
  • Skip-taps
L5Lecture

Decoder · U-Net

Symmetric expanding path: transposed convolutions, concatenated skip connections, double-conv blocks, dropout, and the 1×1 softmax head that paints the per-pixel damage map.

  • UpConv
  • Skip-concat
  • Pixel-accurate
L6Lecture

Ablation Study

What an ablation study is, why we need it, and a chart-by-chart walkthrough of the four-architecture comparison on xBD — explained for undergraduates.

  • What-if analysis
  • Bar charts
  • Per-class F1
L7Lecture

Hyper-parameters

Learning rate, batch size, optimiser, weight decay, dropout, schedulers, and the exact settings used for the xBD project — with figures, equations, and a tuning workflow.

  • η, B, AdamW
  • Schedulers
  • Tuning cheat-sheet
PLAN

Suggested 3-week reading schedule

A paced plan that assumes ~45–60 minutes of focused reading per day. Each row pairs a lecture with a concrete "you can do this after" goal. Stretch it to 6 weeks if you also code along in the FastAPI track.

WeekDaysLectureTimeBy the end you can…
Week 1Day 1–2L0 · Project Overview90 minUnderstand the end-to-end pipeline, the xBD dataset, and what an ATC-20 placard is.
Week 1Day 3–5L1 · Vision Transformers2 × 60 minMaster patches, self-attention, multi-head, and why residuals defeat vanishing gradients.
Week 2Day 1–2L2 · Focal, Dice & Class Weights75 minExplain why plain cross-entropy fails on xBD and how each loss component fixes a different failure mode.
Week 2Day 3–4L3 · Siamese Networks75 minReproduce the shared-encoder + difference-block diagram from memory.
Week 2Day 5L4 · Encoder · ResNet-5060 minWalk the stem → 4 bottleneck stages → skip-taps without checking notes.
Week 3Day 1L5 · Decoder · U-Net60 minTrace how a 16×16 deep feature becomes a 512×512 damage map via skip-concat + up-conv.
Week 3Day 2–3L6 · Ablation Study75 minRank the four components by Δ-F1 and justify the order in writing.
Week 3Day 4–5Recap + Quiz90 minScore ≥ 80% on the 50-question quiz; revisit any lecture where you scored < 70% on its questions.
How to read each lecture
  1. Read the "In plain English" primer at the top.
  2. Skim every figure caption before reading the body.
  3. Re-derive at least one equation on paper.
  4. Open the related quiz questions and answer from memory.
Total time budget

10 hours of reading + 3 hours of quiz/review across three weeks. Realistic for a student carrying a normal course load.

If you only have one week

Read L0 → L2 → L3 → L6 in that order. You'll know the pipeline, why the losses matter, the change-detection trick, and which components carry the model's accuracy.